Search

Two-dimensional (2D) semiconductors have attracted tremendous interest as natural passivation and atomically thin channels could facilitate continued transistor scaling. However, air-stable 2D semiconductors with high performance are quite elusive. Recently, an extremely-air-stable ${\mathrm{Mo}\mathrm{Si}}_{2}{\mathrm{N}}_{4}$ monolayer was successfully fabricated [Hong et al., Science 369, 670 (2020)]. To further reveal its potential application in sub-5-nm metal-oxide-semiconductor field-effect transistors (MOSFETs), there is an urgent need to develop integrated circuits. Here, we report first-principles quantum-transport simulations on the performance limits of n- and p-type sub-5-nm monolayer (ML) ${\mathrm{Mo}\mathrm{Si}}_{2}{\mathrm{N}}_{4}$ MOSFETs. We find that the on-state current in the ${\mathrm{Mo}\mathrm{Si}}_{2}{\mathrm{N}}_{4}$ MOSFETs can be effectively manipulated by the length of gate and underlap, as well as the doping concentration. Very strikingly, we also find that for the n-type devices the optimized on-state currents can reach up to 1390 and 1025 \textmu{}A/\textmu{}m for high-performance and low-power (LP) applications, respectively, both of which satisfy the International Technology Roadmap for Semiconductors (ITRS) requirements. The optimized on-state current can meet the LP application (348 \textmu{}A/\textmu{}m) for p-type devices. Finally, we find that the ${\mathrm{Mo}\mathrm{Si}}_{2}{\mathrm{N}}_{4}$ MOSFETs have an ultralow subthreshold swing and power-delay product, which have the potential to realize high-speed and low-power consumption devices. Our results show that ${\mathrm{Mo}\mathrm{Si}}_{2}{\mathrm{N}}_{4}$ is an ideal 2D channel material for future competitive ultrascaled devices.

We develop a new position-resolved band-unfolding method based on the density functional theory to clarify the single-electron energy spectrum of $(3\ifmmode\times\else\texttimes\fi{}3)$ silicene on $\mathrm{Ag}(111)$ substrate. The position-resolved scheme enables us to clarify each contribution from each spatial region to the single-electron spectrum, which facilitates the chemical identification of each electron state. We find interface states which are distributed in the region of silicene and top two layers of the Ag substrate near the Fermi level and also below the Fermi level. The states are unique in silicene on a substrate in the sense that they are mixtures of Si and Ag orbitals. The obtained electronic structure near the Fermi level is interesting, featuring a hyperbolic-paraboloid-shaped energy band which leads to 12 Dirac-like cones at the boundary of the primitive Brillouin zone of $\mathrm{Ag}(111)$. Characteristics of measured photoemission spectra are satisfactorily explained by the obtained unfolded bands.

Despite the recent progress on two-dimensional multilayer materials (2DMM) with weak interlayer interactions, the investigation on 2DMM with strong interlayer interactions is far from its sufficiency. Here we report on first-principles calculations that clarify the structural evolution and optoelectronic properties of such a 2DMM, multilayer silicene. With our newly developed global optimization algorithm, we discover the existence of rich dynamically stable multilayer silicene phases, the stability of which is closely related to the extent of sp3 hybridization that can be evaluated by the average bonds and effective bond angles. The stable Si(111) surface structures are obtained when the silicene thickness gets up to four, showing the critical thickness for the structural evolution. We also find that the multilayer silicene with pi-bonded surfaces present outstanding optoelectronic properties for the solar cells and optical fiber communications due to the incorporation of sp2-type bonds in the sp3-type bonds dominated system. This study is helpful to complete the picture of structure and related property evolution of 2DMM with strong interlayer interactions., 4 figures

We report on total-energy electronic-structure calculations in the density-functional theory performed for both monolayer and bilayer silicene on Ag(111) surfaces. The rt3 x rt3 structure observed experimentally and argued to be the monolayer silicene in the past [Chen et al., Phys. Rev. Lett. 110, 085504 (2013)] is identified as the bilayer silicene on the Ag(111) surface. The identification is based on our accurate density-functional calculations in which three approximations, the local density approximation, the generalized-gradient approximation, and the van-der-Waals-density-functional approximation, to the exchange-correlation energy have been carefully examined. We find that the structural tristability exists for the rt3 x rt3 bilayer silicene. The calculated energy barriers among the three stable structures are in the range of 7 - 9 meV per Si atom, indicating possible flip-flop motions among the three. We have found that the flip-flop motion between the two of the three structures produces the honeycomb structure in the STM images, whereas the motion among the three does the 1 x 1 structure. We have found that the electron states which effectively follow Dirac equation in the freestanding silicene couple with the substrate Ag orbitals due to the bond formation, and shift downwards deep in the valence bands. This feature is common to all the stable or metastable silicene layer on the Ag(111) substrate., 7 pages, 4 figures

silicene in current experiments, we find several stable and metastable structures with the 4 × 4, √ 13 × √ 13, and 2 √ 3 × 2 √ 3 periodicities with respect to the 1 × 1 Ag(111) lateral cell within the total-energy difference of 70meVperSiatom.Thosestablestructuresshowtheexcellentagreementwiththescanningtunnelingmicroscopy measurement in their structural characteristics. The metastable structures with comparable total energies await experimental observations. In all the stable and metastable structures, the silicene is buckled substantially so that the π state rehybridizes with the σ state, leading to the π + state, and then the linear energy dispersion peculiar to the Dirac electrons disappears in several cases associated with the opening of the energy gap. Moreover, we find that the substantial mixing of the π + state, generated in such a way, with the states of the Ag atoms in the substrate converts the π + state to the mixed π + state and thus makes the state shift downwards or upwards, eventually annihilating Dirac electrons near the Fermi level. The absence of Dirac electrons caused in this way is found to be common to all the stable and metastable structures of the silicene on the Ag(111) substrates. We also find that the interaction between the π + and the substrate orbitals should be weak enough to preserve Dirac electrons and at the same time be sizable to keep the system stable. We then propose two specific substrates as good candidates for the silicene with Dirac electrons, i.e., hexagonal BN and the hydrogen-processed Si(111) surface. We clarify that the silicene on those substrates are stable enough with the binding energy comparable to or twice that of the graphite and preserve Dirac electrons near the Fermi level.

We study the effect of SiC substrate on thermal conductivity of epitaxial graphene nanoribbons (GNRs) using the nonequilibrium molecular dynamics method. We show that the substrate has strong interaction with single-layer GNRs during the thermal transport, which largely reduces the thermal conductivity. The thermal conductivity characteristics of suspended GNRs are well preserved in the second GNR layers of bilayer GNR, which has a weak van der Waals interaction with the underlying structures. The out-of-plane phonon mode is found to play a critical role on the thermal conductivity variation of the second GNR layer induced by the underlying structures.

The Stokes and anti-Stokes Raman spectra of a multiwalled carbon nanotube (MWNT) sample are studied here by four excitation energies and the observed Raman modes are assigned based on the double resonance Raman effect andthe previous results in graphite whiskers. There exists frequency discrepancy between Stokes and anti-Stokes lines (FDSA) of many Raman modes in MWNT's and the discrepancy values are strongly dependent on the excitation energy, in which the FDSA value of the D' mode even changes from a positive value (9 cm - 1 , 1.58 eV) to a negative value (-11 cm - 1 , 2.54 eV). The laser-energy dependence of the FDSA values of some modes in MWNT's is attributed to the nonlinear frequency dependence of Stokes and anti-Stokes Raman lines of these modes on the excitation energy. Raman results and the theoretical analysis of the intravalley and intervalley double resonance processes of Stokes and anti-Stokes Raman scatterings both show that the frequency of an anti-Stokes peak excited by e L is equal to that of the corresponding Stokes peak excited by a laser excitation of e L +∞ω S where ∞ω S is the phonon energy of the Raman mode. Stokes and anti-Stokes double-resonance Raman scatterings have been used to probe the phonon dispersion relations of graphite. The Raman data of the well-known disorder-induced D mode are in good agreement with the theoretical results.